1. Field of the Invention
The present invention generally relates to microelectronic chip design, and more particularly to a method of modeling a logic cell for circuit level models.
2. Description of the Related Art
Generally, integrated circuits are electrical circuits comprised of transistors, resistors, capacitors, and other electrical components on a single semiconductor chip in which the components are interconnected to perform a given function. A circuit designer designs an integrated circuit by creating a circuit schematic indicating the electrical components and their corresponding interconnections. Moreover, designs are simulated by computer-aided design programs to verify the functionality and to ensure that performance goals are satisfied.
The layout of an integrated circuit contains parasitic resistances and capacitances from the interconnections and devices. The values of these so-called parasitics are functions of the process parameters, shape and dimensions of a particular geometry, and relationship of a particular geometry to other geometries. These parasitics affect the performance and possibly the functionality of an integrated circuit. Consequently, during the design phase of an integrated circuit, these parasitics are extracted from a layout and taken into consideration during circuit simulation.
Integrated circuits which are fabricated using deep sub-micron processes are becoming mainstream, but are complex. One of the challenges faced by today""s deep sub-micron integrated circuit designers is the issue of parasitic effects of passive interconnect ions. Deep sub-micron integrated circuit designers are recognizing that these effects cannot be ignored, otherwise the design can fail. These effects also play an important role in timing, power, reliability, as well as noise performance. In order to take parasitic effects of passive interconnects into consideration in post layout analysis for timing, power, etc., it is necessary to create electrical models for the physical connections present between the various devices in a deep sub-micron integrated circuit design, a process known as parasitic extraction. As such, multiple parasitic extractions performed repeatedly at different points in time to create multiple electrical models for different views of the design are often required.
General designs are created using hierarchical methodologies. Here, small circuits are designed and evaluated, and then connected to larger functions until the full chip is assembled. Each element of the hierarchy is called a cell (logic cell) and the basic circuit functions, which include devices and interconnects, are called leaf-cells.
The problem associated with most conventional extraction programs is that they work in a flat manner. That is, the conventional extraction programs ignore the design hierarchy. Some packages have the option to use design hierarchy without a check and simply ignore placement dependent changes. These conventional systems make the user responsible for the inaccuracies.
More sophisticated solutions use an active radius around each wire and check whether the environment within this radius is constant. Variations outside of this radius are considered small and ignored. However, because these solutions are often rigid in their application, this leaves open the possibility that some variations important in the hierarchical design are ignored, thus resulting in an inaccurate manner in which to model a logic cell. Therefore, there remains a need for a new and improved process and system of modeling a logic cell using parasitic extraction techniques in a worst-case environment for the logic cell.
In view of the foregoing and other problems, disadvantages, and drawbacks of the conventional modeling techniques for logic cells using hierarchical extraction processes, the present invention has been devised, and it is an object of the present invention to provide a system and method for modeling a logic cell.
In order to attain the object suggested above, there is provided, according to one aspect of the invention, a method of extracting circuit characteristics from a circuit design that extracts first cell characteristics from a portion of the circuit design using a first set of environmental conditions. Next, the invention extracts second cell characteristics from the portion of the circuit design using a different, second set of environmental conditions. The invention determines the difference between the first cell characteristics and the second cell characteristics. The invention compares this difference to a predetermined standard. The invention labels the portion of the circuit design as freely placeable within any area of said circuit design if the difference is less than the predetermined standard. Next, the invention replaces the portion of the circuit with a placeholder cell if the portion of the circuit design is freely placeable. The invention forms the placeholder cell in a process that includes shorting all conductors in the portion of the circuit design to a ground node. The invention then merges all the conductors in a given level of the portion of the circuit design. Next, the invention removes all the conductors that are covered by overlying conductors from the portion of the circuit design. The invention merges the conductors outside the portion of the circuit design that are within a predetermined distance to the circuit design to conductors within the portion of the circuit design. The circuit characteristics comprise at least one of capacitance, impedance, power, and resistance.
The first environmental conditions represent the best environment and the second environmental conditions represent the worst environment. The best environment has the minimum amount of wiring adjacent the portion of the circuit and the worst environment has the maximum amount of wiring adjacent the portion of the circuit. The invention calculates average cell characteristics from the portion of the circuit design for the placeholder cell.
To decouple the full (schematic and parasitic) model of one cell from its environment both the impact of the environment on the cell and the impact of the cell on the environment must be considered. Generally, according to the present invention, the amount of variation a cell can tolerate must be determined. Then, a calculation of how much variation different placements can cause is performed. A decision of whether the cell is freely placeable is made based on the level of variation for that cell.
Designed blocks of circuitry or cells are preferably reused in different environments, with different wires running over, next to, or even through the previously characterized design. If all these different environments do not change the characteristic of the description of this cell significantly, the cell can be placed anywhere on the die without timing model changes, and it is considered freely placeable. When the cell is found to be freely placeable the cell undergoes parasitic extraction in an xe2x80x9caveragexe2x80x9d environment (centered between the worst and best environment standards) and the timing model for the cell is created. The model can be stored for use with the final assembly of the part with all leaf cells and interconnect cells, or it can be characterized and simplified, when a final detailed assembly is not required. The next step the invention takes is characterization of the parent cell without the just characterized cell being present. For this step the invention replaces the cell with a simplified model leaf cell, which just contains the important elements for characterization of the interconnect cell. This leaf cell model is sometimes called a placeholder. This highly simplified placeholder design contains just the features which can impact the extraction of the interconnect, or more generally the environment of the leaf cell, so that internal extraction calculations internal to the leaf cell itself can be dropped out of further extraction processes. This placeholder is an automatically created design, which will never be released, and is just built to enable efficient extraction.
The invention makes this placeholder a virtual design, without ever actually creating the shapes and nodes the placeholder contains. The most efficient creation of the leaf cell depends on the programming environment. The placeholder is created by first connecting all nodes within the leaf cell together to one single unique node of this cell. This brings all shapes onto the same potential, still adjustable together for sensitivity studies from the outside. This method is accurate when the potential changes within the cell can be neglected.
In addition, the invention merges all the shapes which are overlapping within the leaf cell. This step reduces the number of edges within the leaf cell, thus simplifying the extraction from outside. Also, shapes which are close but not overlapping can be merged without much loss of accuracy. For this step, a closeness threshold is used. This step relies upon the fact that the fringe capacitance each shape makes the shape look larger from the outside. All hidden shapes are also removed, as they cannot be seen from outside and hence do not impact the extraction of the parent cell. Then, this highly simplified placeholder design is used to represent the cell to minimize the impact of the environment extraction which is tied to its own node. The extraction of the environment cell or parent cell is preformed with the placeholder present. This makes this extraction step much more accurate than just dropping the leaf cell out of extraction as done in many products.
The degree of simplification affects the extraction runtime improvements. Naturally, the runtime improvement depends on the degree the placeholder could be simplified in comparison to the full model. Moreover, ports are created at the contact point to the simplified placeholder cell model. Then, the environment is extracted with the placeholder using changed capacitance formulas for any capacitance going to the placeholder node. The new formulas included in the tolerance are the impact of possible active nodes in the virtual cell. For all practical purposes, this node will be at ground. But to check the sensitivity, it can be of benefit to separate the two nodes. To measure the capacitance connected with ground, for ground bounce tests, it is important to have the nodes separated. Finally, the previously extracted cell model is added to the extracted net list. For every instance of the cell, a placeholder is added. So it is an advantage to have the placeholder significantly simpler than the real cell.
The present invention provides a powerful method which allows the ability to reuse characterized cells multiple times and in multiple instances, wherein different instances of the same cells are in different environments in the chip assembly. The extraction process calculates the resistances and capacitances (RC) of parasitic elements including all connecting conductors and wires. Because capacitances of the interconnect depend of the environment, conventional systems have not been able to reuse a once extracted cell over many instances. Thus, the present invention describes a method to rigorously evaluate the cases in which a hierarchy can be used.